June 2004: Part 6
Previous  Top  Next

24/6/2004
Window coating design with CODE
Today we start serious work on the coating design for architectural and automotive glass. From M.Theiss Hard- and Software we have received a starting configuration that we are going to use in the future. We save this configuration in the database (subfolder 'Software configurations' and there the folder 'CODE') using the name 'window_design.wcd'.

The configuration computes the following optical spectra:
·Reflectance in the wide spectral range 200 ... 50000 nm using 2000 data points  
·Transmittance in the wide spectral range 200 ... 50000 nm using 2000 data points  
·Absorption (=1-R-T) in the wide spectral range 200 ... 50000 nm using 2000 data points  
·Backside reflectance in the wide spectral range 200 ... 50000 nm using 2000 data points  
·Reflectance in the visible (380 ... 780 nm) using 81 data points  
·Transmittance in the visible (380 ... 780 nm) using 81 data points  
·Absorption (=1-R-T) in the visible (380 ... 780 nm) using 81 data points  
·Backside reflectance in the visible (380 ... 780 nm) using 81 data points  

Based on these spectra several technical data are computed. The computed spectra and data are displayed in several views in the main window.
The numerical values are shown in the first view which is displayed when you load the configuration. For a 4 mm pane (Float A) the following values are achieved:

clip0043

Clicking the >
symbol in the lower right corner you can switch to the next view which shows the spectra in the visible:

clip0044

The spectra computed in the large spectral range are used to compute solar reflectance and transmittance, emissivities, U and g. The next view shows the spectra on logarithmic wavelength scales:

clip0045

Here we notice that the mid infrared spectra (2500 ... 50000 nm wavelength) are not correct! The transmittance of the glass pane should be 0, and the reflectance spectrum should show typical features of strong vibrational modes. We will have to fix this problem in order to trust emissivity computations.

However, for the moment we continue with the inspection of the CODE configuration for window design. The program tries to visualize the appearance of the coating in two additional views. One shows the colors

clip0046

and the last one a picture of the 'Virtual office tower'. For every window of the building CODE computes the angle of incidence, the reflectance spectrum for this angle, its color and the corresponding visual appearance:

clip0047

Well, that's nice, but before we really use this configuration for non-trivial cases we would like to repair the mid-infrared errors of the computed spectra.